基于希尔伯特——黄变换的GIS超声波检测异常信号诊断

Fault Diagnosis of Gas Insulated Switchgear (GIS) Based on Hilbert-Huang Transform (HHT) of On-line Ultrasonic Detection Signals

文中提出了一种气体绝缘组合电器(gas insulated switchgear,GIS)超声波检测异常信号的时频特征计算与诊断方法。首先运用希尔伯特-黄变换(Hilbert-Huang transform,HHT)得到描述超声波信号包络线的时频特征的Hilbert幅值谱。在此基础上,给出了离散条件下Hilbert边际谱与平稳性指数的计算方法,分别表示信号中各频率成分的总强度和平稳性。然后将Hilbert边际谱与平稳性指数结合起来分析,可以直观读取信号的时频特征,进而推断出异常气室的故障原因。3个现场实测案例的研究表明,时频特征分析方法可以有效反映电晕放电、悬浮电位放电和机械振动等3种不同故障对应的超声波信号的时频特征,对提升GIS超声波检测结果的分析诊断水平有着重要的指导意义。

A fault diagnosis approach for gas insulated switchgear(GIS)is proposed based on the time-frequency characteristic of on-line ultrasonic detection signal.Firstly,the Hilbert amplitude spectrum of the signal envelopes is obtained by using Hilbert-Huang transform(HHT).Then,the Hilbert marginal spectrum and the smoothness index are derived,which express the total intensity and smoothness of different frequency components in the ultrasonic signal,respectively.The Hilbert marginal spectrum is combined with the smoothness index to intuitively illustrate the time-frequency characteristics of the ultrasonic signal,hence the abnormal condition in the GIS chamber could be inferred.Three practical cases show that the present approach can identify the time-frequency characteristics of the on-line ultrasonic detection signals relating to corona partial discharge,floating potential partial discharge and mechanical vibration.